Sandglasset：一种用于时域语音分离的轻型多粒度自关注网络
Sandglasset: A Light Multi-Granularity Self-attentive Network For Time-Domain Speech Separation
摘要：
领先的单通道语音分离（SS）模型之一是基于具有双路径分段技术的TasNet，其中每个段的大小在所有层中保持不变。相比之下，我们的主要发现是，多粒度的功能是必不可少的，以提高上下文建模和计算效率。我们引入了一种具有新颖沙漏形状的自我关注网络，即Sandglasset，它以显著较小的模型大小和计算成本提高了最先进的（SOTA）SS性能。沿着Sandglasset内部的每个块向前，特征的时间粒度逐渐变得更粗，直到达到网络块的一半，然后朝着原始信号水平逐渐变得更细。我们还揭示了具有相同粒度的特征之间的残余连接对于通过瓶颈层后保留信息是至关重要的。实验表明，我们的Sandglasset只有2.3M的参数已经取得了最好的结果在两个基准SS数据集-WSJ 0 - 2 mix和WSJ 0 - 3 mix，其中SI-SNRi分数已分别提高了绝对0.8 dB和2.4 dB，与以前的SOTA结果相比。
英文摘要：
One of the leading single-channel speech separation (SS) models is based on a TasNet with a dual-path segmentation technique, where the size of each segment remains unchanged throughout all layers. In contrast, our key finding is that multi-granularity features are essential for enhancing contextual modeling and computational efficiency. We introduce a self-attentive network with a novel sandglass-shape, namely Sandglasset, which advances the state-of-the-art (SOTA) SS performance at significantly smaller model size and computational cost. Forward along each block inside Sandglasset, the temporal granularity of the features gradually becomes coarser until reaching half of the network blocks, and then successively turns finer towards the raw signal level. We also unfold that residual connections between features with the same granularity are critical for preserving information after passing through the bottleneck layer. Experiments show our Sandglasset with only 2.3M parameters has achieved the best results on two benchmark SS datasets -- WSJ0-2mix and WSJ0-3mix, where the SI-SNRi scores have been improved by absolute 0.8 dB and 2.4 dB, respectively, comparing to the prior SOTA results.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
